Wednesday, December 14, 2011

After Careful Consideration

Recently I was asked my opinion about the future of private clouds. It's actually something that I spend a good deal of time thinking about, and I had an answer ready. As much as I'd like to think that my response was insightful, that question is one that's best posed to my colleagues who posed the question.

And if any of them are reading this now, I'd like to qualify my answer.

At the time I said something to the effect of, "as long as there are computing resources within an enterprise, it only makes sense to attempt to leverage the unused capacity."  I have no problem with my answer--at least in the short-term.

If you think about it, desktop machines are particularly powerful (and certainly getting more powerful all the time--Moore's Law, incredibly simplified), and it certainly makes sense to try to harness all of that unused computing capacity.  Years ago, my employer used to quote a statistic that typical machines had utilization numbers in the single digits. If you think about it, how often do you really harness the complete capacity of your computer anyway?  Unless you're constantly running complex calculations, it's unlikely (gaming may, but checking email and performing other routine business functions don't typically consume a lot of CPU power). As computers become more powerful, your utilization only drops further. My last computer had 16 CPU's--there are so many things it could've been doing in addition to everything I needed it to do.

Once you look at this in a business context, you can start to see the need to harness this capacity... and that's exactly what a private cloud would do--give you the ability to use these untapped resources without having to purchase additional machines to address the needs of applications that really demand a lot of power. It's particularly valuable because by increasing the usage of what you already have, you can eliminate the cost of new purchases... and you can reduce the high costs of housing and maintaining your current server farm by essentially doing more (compute) with less (resources)-- another benefit of increasing utilization.  Doing more with less also allows you to reduce what you currently have.

It's not really difficult to see the need for private clouds. Even without a dedicated datacenter, every desk typically has a computer on it... it is the 21st century after all.

But here's what I've been thinking about since my discussion last week, and the reason I want to qualify my answer: I was thinking in current terms, and hadn't considered the reality of the ever-nearing future.

Why does every desk need a computer?  It's the 21st century?  I think quite a few people could get away--and benefit--from not having a computer. It shouldn't be that shocking. Ever hear of the cloud? The PUBLIC cloud?

Most of what you already do is available there--email and the Microsoft Office products--among countless other offerings. If you can access what you need from wherever you are, why do you need to be chained to your desk? As this awareness continues to grow, the demand for private cloud will shrink... and the datacenter will go the way of other things that at some point seemed impossible to do without: the home telephone, the VCR, film, being the same room to conduct business, among them.

So that's my thought: there's a big need for private clouds today, but there won't be tomorrow. The change is as inevitable as technology itself.

To my friends, if you're reading this, that's my final answer.